(Mammalia, Carnivora) in the Miocene of Bulgaria

Mitt. Bayer. Staatsslg. Paläont. hist. Geol. 42 83-101 München, 15. 12. 2002

The first appearance of Dinocrocuta gigantea and Machairodus aphanistus (Mammalia, Carnivora) in the Miocene of Bulgaria

N ikolai Spassov & G eorge D. K oufos With 9 figures

Abstract

Some interesting carnivores from the late Miocene of Bulgaria coming from two different localities are studied in this article. A part of the material was found in the locality of Nessebar, near Bourgas and the rest in the area of Blagoevgrad (SW Bulgaria). The studied material has been determined to Dinocrocuta gigantea and Machairodus aphanistus. Both species were unknown in the Bulgarian late Miocene until now and their presence enlarges our knowledge about their distribution. The age of both localities is also discussed on the basis not only of the carnivores but of the rest fauna too. The localities have been dated to Vallesian without possibility for a more precise age ( MN biozones) because of the scarce material and the lack of sufficient data.

Kurzfassung

Es werden einige interessante Carnivoren aus den Obermiozän von Bulgarien beschrieben, die aus zwei verschiedenen Gebieten kommen. Ein Teil des Materials wurde an der Lokalität Nessebar bei Burgas gefunden, das übrige im Gebiet von Blagoevgrad (SW-Bulgarien). Das untersuchte Material wurde als Dinocrocuta gigantea und Machairodus aphanistus bestimmt. Beide Arten waren im bulgarischen Obermiozän bisher unbekannt und ihre Gegenwart erweitert unsere Kenntnis über ihre Verbreitung. Das Alter beider Fundstellen wird nicht nur auf der Basis der Carnivoren, sondern auch der Gesamtfauna diskutiert. Beide wurden ins Vallesium eingestuft, da das spärliche Material und das Fehlen weiterer Daten eine genauere Einstufung (MN Biozonen) nicht erlauben.

*) Addresses of the authors. N. Spassov , National Museum of Natural History, Blvd.T zar Osvoboditel1 , 1000 Sofia, Bulgaria. G. D. K oufos , Aristotle University of Thessaloniki, Laboratory of Geology and Palaeontology, 54006 Thessaloniki, Greece, e-mail: koufos(3'geo.auth.gr

83 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

Aivd <>l llit' most prol\ihlf plat e of the Bldgtx'vgrad lot dlilv Fig. I. Map of Bulgaria indicating the geographic position of the localities.

1. Introduction

The late Miocene localities of Bulgaria are quite numerous, while a great amount of fossils have been collected (Bakalov & N ikolov 1962; N ikolov 1985; Spassov 2000). The material is stored in the National Museum of Natural History (NMNH) of Sofia, its branch in Assenovgrad and in the University of Sofia. Among the collections of the NMNII one of the authors (N. S.) found some remains of carnivores from two different localities. One tooth of a large percrocutid and another of a machairodont come from the locality of Nessebar (Fig. 1). The specimens are labeled by 1. Nikolov asMa.chairod.ns sp. Some fragments of teeth and one lower carnassial of a large percrocutid are also found with a label referring“Percrocuta senyiireki“ and Blagoevgrad area as their locality. The sediment’s remains on the teeth indicate that they come from coal-bearing layers. Probably the teeth were brought to I. Nikolov by some industrial research geologists or the miners. Since now, the large percrocutids were unknown in the Miocene of Bulgaria and very rare in Europe. In the Balkan Peninsula they are known from three localities.Percrocuta miocemca is known from the middle Miocene locality of Prebreza, in Yugoslavia(P avlovic & T henius 1965). Two large forms are also referred from the late Miocene of Axios valley (Macedonia, Greece), Dinocrocuta gigantea from the Vallesian locality of “Pentalophos 1“ andDinocrocuta salomcae from the PVallesian locality of Diavata (K oufos 1995). The genus Dinocrocuta is also referred from the PTurolian localities of Los Aljezares and Adernuz, Spain by few remains (Soria 1980). Very fragmentary remains, probably belonging to a relatively small Dinocrocuta, are recently found in the early Turolian locality of Dorn-Dürkheim 1, Germany (Morlo 1997).

84 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

The Miocene Machairodus finds are also very rare in Bulgaria. A lower carnassial of “Machairodus schlosseri“ (= Paramacbairodus orientalis) and a Machairodus humerus are re ferred from Kalimanci district (Bakalov & N ikolov 1962). An undescribed skull associated with the mandible of Machairodus giganteus is also known from the Turolian locality of Hadjidimovo and stored in the Palaeontological Museum of Assenovgrad(K ovachev pers. comm.; Spassov in press). The studied specimen was found in the locality of Nessebar (Fig. 1) and the attribution to Machairodus aphanistus certifies the presence of this species in Bulgaria.

2. Localities

The locality of Nessebar (old Turkish cemetery) is situated northeastern to the city of Bourgas across the Bulgarian seashores of the Black Sea (Fig. 1). The initial collection from this locality was described by Bakalov & N ikolov (1962), while later revised by N ikolov (1985). According to the later author the deposits consist of sandstones and limestones. The few traces of the sediments on the studied specimens suggest that they possibly come from the sandstone level. The determined mammalian fauna from Nessebar suggests Sarmatian (= Vallesian) age, while the mollusks indicate a middle Sarmatian age (N ikolov 1985). The exact site of the Blagoevgrad material is unknown. Taking in mind, the traces of lignite on the fossils and their taxonomic status, the studied remains originate from the Neogene lignitic deposits of the area. The lignitic deposits outcrop near the villages of Oranovo and Arnautska Mahala, about 10 km south of Blagoevgrad (Fig. 1). The age of the deposits is questionable yet, most probably Sarmatian-Maeotian, i.e. roughly corresponding to Vallesian (Marinova 1993).

3. Systematic Palaeontology

Family Percrocutidae W erdelin & Solounias 1991

Genus Dinocrocuta Schmidt -K ittler 1976

Dinocrocuta g i g a n t(S echlosser a 1903) (Figs 2,3)

Localities: Nessebar (old Turkish cemetery), Bourgas area, SE Bulgaria; Blagoevgrad vicinity (Oranovo - Arnautska Mahala region?), SW Bulgaria. Age: Vallesian, late Miocene. Material: Nessebar: Left upper carnassial, FM 1504. Blagoevgrad: Left P, FM 1500; fragment of the right upper canine, FM 1502; fragment of the left lower canine, FM 1501; right lower carnassial, FM 1503. All specimens from Blagoevgrad probably belong to one individual. Measurements (in mm): P4 FM 1504: Length=50.8; Length paracone+metacone = 33.9; Breadth at the protocone = (25); Breadth at the parastyle = 16.5 P FM 1500: Mesiodistal diameter = 18.0 Buccolingual diameter = 13.5 Cs FM 1502: Mesiodistal diameter = 29.0 Buccolingual diameter = - Mj FM 1503: Length = 37.8; Breadth = 18.2; Length of trigonid = 34.3

85 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

Description. P. It is robust, canine-like (Fig. 3a), flattened laterally and it has a very large contact surface with I2. The tip of the cusp is curved distally and preserves a labial and lingual crest. The latter contacts a triangular, well pronounced lingual cingulum. Cs. The available specimen preserves the upper part of the canine having a weak wearing facet for the lower canine on its frontal surface (Fig. 3b). P4. This unique tooth from Nessebar is totally unworn and belongs to a juvenile-subadult individual (Fig. 2). Before the preparation of the tooth part of the maxillary the bone was in connection with the basal half part of the tooth, showing that it was still not completely erupted. The tooth is long and narrow labio-lingually. The lingual surface of the cusp is more rounded than the labial one. The parastyle is relatively reduced, strongly pointed and assymetrical with strongly inclined and much longer mesial surface (Fig. 2c). A well pronounced vertical crest is present in its lingual and mesial surface and a weaker third one between them (Fig. 2c). The enamel crust of the protocone was damaged but we can judge about its form by the calciferous stuffing which forms a “cast“ of the protocone. The protocone is reduced and directed obliquely in front and upwards. Its transverse diameter is at least V3 shorter than the transverse diameter of the parastyle. The protocone seems to lack a real tip and it is like a mesiolingual projection of the base of parastyle and paracone (Fig. 2 a,c). The mesial border of the protocone does not exceed the mesial surface of the parastyle, while its distal border reaches the contact between the parastyle and the paracone. A well expressed transversal groove separates the parastyle from the paracone. The paracone is longer, broader and much higher than the parastyle. It preserves a well pronounced crest on its mesial surface, a trace of such a crest from the top on the labio-distal surface, and a sharp edge on its distal surface. The metastyle is slightly higher than the parastyle. It is long, evidently longer than the paracone and it is separated from this cusp by a deep and narrow groove. A slight, groove-like, depression is marked from the top of the labial surface of the metastyle separating this in two (anterior and posterior) parts. In occlusal and lateral view the paracone and metastyle edges form a wide angle, indicating high cutting specialization. A weakly pronounced cingulum is present in the lingual surface of the paracone-metastyle complex. No cingulum exists on the labial surface of the tooth. C. The specimen lacks the root (Fig. 3c). A clear wearing facet is distinguished in its distolabial surface. Mr It belongs to an adult individual with strong wearing facet in the para- and protoconid (Figs 3d-f). The lower carnassial has a very strong mesial root and a strong derived trigonid; its vestibular surface is slightly convex and the lingual one slightly concave. The paraconid is longer and broader than the protoconid but the latter is evidently higher than the former. A crest runs across its lingual surface from the tip to the mid-height (Fig. 3d). The paraconid mesial surface is oblique and has a cutting edge, partially worn. Traces of a strong cingulum exist on its vestibular surface. The para- and protoconid are separated by a shallow groove. In occclusal and lateral view the cutting edges form a wide angle, indicating a derived carnivorous structure. The metaconid is absent. The talonid is relativelly small and well separated from the trigonid. The entoconid exists and seems to be higher than the hypoconid but this is probably due to the stronger attrition of the vestibular side of the tooth. A small distal cingulum is present behind the two mentioned cuspids of the talonid.

Comparison and discussion:

The studied upper carnassial differs from the large Felidae (Machairodus) because it is more robust, lacks a preparastyle and has a relatively shorter metastyle. The lower carnassial differs

86 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

b Fig. 2. Dinocrocuta gigantea, Nessebar, Bulgaria, Vallesian. Left upper carnassial, FM 1504; a. lingual, b. labial, and c. occlusal view. in having a longer and broader paraconid, a more developed and wider (not cutting) talonid and an expressed cingulum. The studied P4 differs from the late Miocene Barbourofelispiveteaui from Turkey (G eraads & G ulec 1997) by the absence of the preparastyle, the presence of a protocone, the larger parastyle and the larger dimensions. The lower carnassial is significantly reduced in B. piveteaui. It differs from Adcrocuta by: a) the larger size, b) the strongly back wardly inclined mesial surface of the parastyle, c) the relatively longer parastyle with stronger lingual crests, d) the more distally situated protocone, e) the less curved laterally distal part of the metastyle, and f) the relatively reduced talonid with the absolute lack of the metaconid. On the contrary their large size and morphology indicate that they belong to the large percrocutids. The percrocutids were referred to the family Hyaenidae for a long time (K urt Ln 1957; Schmidt -K itti.er 1976;H owf.ll & Petter 1985). Recently, they were separated from Hyaenidae

87 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

as a phylogenetically distinct group constituting the family Percrocutidae, (Werdelin & Solounias 1991;W erdelin 1996). One of the main reasons for that is the different structure of the deciduous teeth, which is close to that of Stenoplesictis and Felidae (C hen & Schmidt - K ittler 1983). The scarce material makes the taxonomy of the family difficult and this reflects to the large synonymies. Barishnikov & A verianov (1995) note that the dp4 of Proteles have the same position of the metaconid as in Percrocutidae and those taxa should be included in the same family. As far as Protelidae G eoi froy 1851 is a senior synonym this name must have the priority. Indeed, if we assume that this feature is homologous in Percrocutidae andProteles and not homoplasic for the noted taxa, Protelidae should be the valid name. The raised problem could not be considered finally solved. According toW erdelin (1996) the family Percrocutidae includes the genera Percrocuta, Dinocrocuta, and possibly Allobyaena. The members of the genus Percrocuta have definitely smaller dental dimensions than the studied remains from Bulgaria. Allobyaena is obviously smaller, having a robust protocone on P4 and a vestigial metaconid on (H owell & Petter 1985). The main characters of Dinocrocuta as given by the various authors with some new ones added are summarized in K oufos (1995). Several species of Dinocrocuta are known and a metrical comparison of the studied material with them is given in Figs 4, 5. In these diagrams the studied teeth are very close to those of D. gigantca but also to some other forms, while they are well separated from D. salonicae, D. algeriensis and D. grandis. So, a morphological comparison is necessary for their clear determination. However, the scarcity of the material from the various localities makes difficult a complete and certain comparison. In most cases the available material is either very few or fragmentary or it belongs to different jaws. The type material of D. gigantea is stored in the Bayerische Staatssammlung für Paläonto logie und Geologie, München (BSP) and comes from an unknown locality of China(S chlosser 1903). One of the authors (G. K.) studied this chinese material and a comparison with the Bulgarian one is given below. The material is very fragmentary, including isolated teeth or pieces of teeth, while a part of the material is missing, i.e. one upper carnassial(S chlosser 1903; pi. II, fig.2). But the illustrations of the last author are nice and allow a comparison. The protocone is strongly reduced like in the studied upper carnassial, while the parastyle seems to be larger in the studied one. The cusps of the chinese specimen seem to be higher but this is probably due to the young age of the studied tooth (it is not fully erupted in order to get its final size). Among the type material there is one complete (BSP 1900 XII 537) and a partial one preserving the posterior half of the tooth (BSP 1900 II 524). Their morphology fits quite well with the studied lower carnassial, except BSP 1900 XII 537 which lacks the entoconid. Concerning the size of the studied material it is very close to that of D. gigantea from China (Figs 4, 5). A mandible ofD. gigantea has been described from the Vallesian locality of „Pentalophos 1 “ (PNT) of Axios valley, Macedonia, Greece(K oufos 1995). The studied M, has a similar size and morphology with the Pentalophos one. The latter has a slightly larger talonid and in the basis of the distal edge of the protoconid there is a small incision. According to one of us (N. S.) this incision indicates the position of the metaconid. A maxilla with P2-P4 of a percrotutid is also known from the locality of Diavata in Axios valley(A ndrews 1918). The specimen was ini tially referred to “Hyaena “ salonicae and later it followed all the taxonomic changes of the percrocutids. However, its characters indicate that it belongs to Dinocrocuta and it is referred as D. salonicae with strong similarities to D. senyiireki and D. grandis (K oufos 1995). The studied P4 is larger than that of D. salonicae and evidently slender; the robusticity index of the former tooth is 49 versus 55.5 in D. salonicae. The most significant difference is the larger protocone ofD. salonicae which is well separated from the parastyle and has well marked tip. A detailed comparison for the distinction ofD. salonicae from D. gigantea is given in K oufos (1995).

88 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

Fig. 3. Dmocrocuta gigantea, Blagoevgrad district, Bulgaria, PVallesian. a. Left 13, FM 1500, labial view, b. Fragment of the right upper canine, FM 1502, lingual view. c. Fragment of the left lower canine, FM 1501, labial view. d-f. Right lower carnassial, FM 1503; d. lingual, e. labial and, f. occlusal view.

89 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

A skull ofD. gigantca was recently described from Hezheng, Gansu, China (Qiu et al. 1988). The morphology of P4 seems to be similar with the studied one, preserving a reduced protocone. Unfortunately, no dimensions are given for certain comparison. However, the measurements from the illustrations suggest that the studied carnassial is slightly smaller than the chinese sample. Some material ofD. gigantea is known from the locality of Varnitsa, Moldavia(L ungu 1978; figs 18, 20). According to his description and drawings the upper carnassial is of slightly larger size than the studied one (Fig. 4) and has a reduced protocone which is a continuity of the base of paracone and parastyle. In this feature it resembles to the studied carnassial but the reduction is stronger in the Moldavian one. Concerning the lower carnassial it is close in size and morphology to the Moldavian material.Dinocrocuta gigantca is also known from the Vallesian localities of Eldari 1 and Natlismtsemeli, Georgia (Now, 2002), as well as from the late Sarmatian (=Vallesian) localities of Buzhor 2, Gritchev and Klimentovichi of Ukraine (Qiu et al. 1999). An interesting percrocutid, referred as D. senyiireki, is known from Yassioren, Turkey (Ozansoy 1965). Some fragmentary material of this species is also referred from the localities of Esme-Akcakoy, Kayadibi and Inonii (Schmidt -K ittlfr 1976). The type material of D. senyiireki includes two mandibular fragments (Ozansoy 1965). The size of M t is very close toD. robusta and D. algeriensis and far from what it is referred as D. gigantea (Fig. 5). However, the upper carnassial referred toD. senyiireki is significantly larger and situated in the group ofD. gigantea (Fig. 4). The comparison of the whole upper dentition from Yassioren indicates that it is closer to D. gigantea and larger than D. algeriensis, D. salonicae and D. robusta. On the other hand the comparison of the lower dentition shows that it is smaller than D. gigantea and closer toD. algeriensis, D. robusta and D. grandis (Koui-os 1995; figs 4, 5). This difference means that either the two specimens do not belong to the same species or there is a great intraspecific size variability. Nevertheless, the Yassioren upper carnassial has a strong, well separated and high pointed protocone like that ofD. salonicae and differs clearly from the studied Bulgarian material. The lower carnassial of D. senyiireki from Yassioren is clearly smaller than that of the studied Mp while the residual metaconid differentiates it from the studied one. D. algeriensis is referred from the Vallesian locality of Bou-Hanifia, Algeria(A rambourg 1959). Besides the smaller size (Fig. 4, 5) it differs from the studied material in having a stronger and

* D. gigantea, Bulgaria

♦ D. gigantaa, China

• D. gigantea, Varnitsa, Moldavia

▲ D. gigantea, Bahe, China □ D. salonicae, Diavata, Greace o D. algariensis, Bou Hanifia o D. senyureki, Vassioran A D. grandis, Siwaliks

Fig. 4. Scatter diagram comparing length / breadth of the upper carnassial in various species ofDinocrocuta.

90 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at well separated protoeone and a more robust P4 (the robusticity index is 63 versus 49 in the studied carnassial). D. algeriensis differs from the studied lower carnassial by the shorter and higher trigonid, the relatively more robust and tricuspid talonid and the presence of a vestigial metaconid. The studied teeth differ approximately in the same features from D. robusta of Kalfa, Moldavia (Lungu 1978). A maxillary fragment with P2-P4 (V. 3093) is known from the Bahe Formation, Shanxi, China and it is referred as Crocuta (Percrocuta) macrodonta and later transferred to D. gigantea (H owell & Pi tter 1985). The shape of the protocone of P4 is closer to that of the studied one, as well to D. gigantea from the Vallesian of Varnitsa, Moldavia. The Chinese and Moldavian forms, however, have more reduced protocones than the studied upper carnassial and may be they represent another form of D. gigantea which can belong to a different subspecies, D. g. macrodonta. A percrocutid from the Siwaliks, India is known asD. grandis. It is remarkably smaller than D. gigantea and the studied teeth. Moreover, the metacone of P4 is shorter relatively to the tooth’s length and the M1 is relatively to P4 shorter than in D. gigantea. Taking into account all the above mentioned comparisons the studied material from Nessebar and Blagoevgrad must be determined to Dinocrocuta gigantea. Dinocrocuta is mainly referred from the Vallesian of Europe but it is also known from the Turolian. It is rarely associated with A dcrocuta and this co-existence is probably not coincidence (H owell & P etter 1985). On the other hand, this rarity is not accepted by W erdelin (1996). In the lower Axios valley, Greece Dinocrocuta was found alone in the Vallesian localities of Pentalophos 1 and Diavata. But, in the slightly younger localities of Ravin de la Pluie, Ravin des Zouaves 1 and Xirochori 1 onlyAdcrocuta is present. Although the two genera have not been found together in any of these localities, they co-exist during Vallesian. Besides this, Dinocrocuta is associated with other small hyaenids; in Pentalophos 1 it was found together with Protictitherium cf. crassum (K oufos 2000a). The genus Adcrocuta is certainly known from late Vallesian of Greece(K oufos 2000a), while its presence is possible in Spain(A lbf.rdi , 1974) and France (H owell & P etter 1985). Adcrocuta gets its great expansion in Turolian and especially in MN 12. During that time the remains of Adcrocuta are numerous in the various circum-Mediterranean localities. Moreover the two genera co-exist in the early Turolian locality of Dorn-Durkheint 1, Germany (Morlo 1997) and Los Aljezares, Spain(S oria 1980). So, it is clear that Dinocrocuta and other percrocutids lived in the Vallesian and they have been replaced by Adcrocuta in the Turolian; however both genera co-exist in the Vallesian/Turolian

D. gigantea, Blagoevgrad

♦ D. gigantea, "Pentalophos 1"

H D. gigantea,China

A D. gigantea, A. Tali

# D.gigantea, Varnitsa

O D. algeriansis, Bou Hanifia

O D. senyuraki, Yassibren

A D. robusta, Kalta

E Dinocrocuta sp., Adamuz

Length Fig. 5. Scatter diagram comparing length / breadth of the lower carnassial in various species of Dinocrocuta.

91 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

transition. It is necessary to refer here that Dinocrocuta re-appears again in the Pliocene of Sahabi with the species D. senyureki (H owei l 1984), if the age of Sahabi fauna is true. The question is why Dinocrocuta and the other percrocutids disappeared and they were replaced by the derived hyaenas like Adcrocuta? The dispersion of the derived hyaenas could be the reason and it is probably due to unknown ecological differences shaping the evolution of the bone-cracking hyaenas and the late percrocutids (Wlrdelin 1996).Adcrocuta resembles to the recent Crocuta in cranial and

C Fig. 6. Machairodus apbanutus, Vallesian, Nessebar, Bulgaria. Lett mandibular fragment with M,, FM 1505; a. labial, b. lingual, and c. occlusal view.

92 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

skeletal morphology, while the analysis of the palaeoenvironment in Adcrocuta bearing localities suggests a mode of life close to the recent. That means a social, active and rapid carnivore, hunting by pursuit and at the same time a bone-cracking scavenger; it inhabited relatively open environments (savanna-like open forests and “forest-steppe“ habitats). Dinocrocuta, on the other hand, has robust and long canines, evolved carnassials and its premolars are bulbous enough to break bones. Thus, this percrocutid attains to one or other extent the trophic specialization of the derived hyaenas and, as it seems, earlier than them. Although it was found in the Turolian of Europe, Asia and Africa too,Dinocrocuta is known mostly from the Vallesian or Vallesian/Turolian localities. Moreover,Dinocrocuta has been found in much less localities than Adcrocuta and the remains are rather scarce everywhere. Having in mind the more wet climate of Vallesian relatively to Turolian in Southeastern Europe and the scarcity of the Dinocrocuta remains one could suppose that the habitats typical for this predator were sylvatic. Two arguments support this statement: the discovery ofD. gigantea remains in the coal layers near Blagoevgrad (Bulgaria) and the more close and wet palaeoenvironment in late Vallesian than in Turolian of Northern Greece(B onis et al. 1999). The noted morpho-functional characteristics ofDinocrocuta, to which we could add the rather large size, together with the taphonomic and chronological distinctions of the localities permit the assumption that Dinocrocuta is a large, solitary supercarnivore with low population size and density compared to Adcrocuta (this is yet another reason for the scarcity of the percrocutid remains).Dinocrocuta should be adapted to kill by ambushing large prey and also to scavenge mostly on carcasses of large animals. The aridity taking place during Turolian gives an advantage to Adcrocuta. In the new palaeoecological conditionsDinocrocuta is “pressed“ between the ecological niches of A. eximia and the large Turolian Macbairodus giganteus (it pursuits and kills large animals, feeding on their carcasses for long period, and inhabits relatively open landscapes). This is a strong reason for the disappearance ofDinocrocuta first from Southeastern and Eastern Europe, where the climatic conditions were more arid, even in the Vallesian. The genus survives longer in isolated regions (Spain) or in the northern more forestrial territories which preserved habitats that were more suitable for its ecology (Germany). There it could co-exist with Adcrocuta in conditions of more diverse environment, occupying different ecological niches. This co-existence is also justified at the end of Vallesian beginning of Turolian representing the transitional time period for the replacement of one genus by the other.

Family Felidae G ray 1821

Genus Macbairodus K aup 1833

Macbairodus aphanist (K aupus 1833) (Fig. 6)

Locality: Nessebar (old turkish cemetery), Bourgas area, SE Bulgaria. Age: Vallesian, late Miocene. Material: Nessebar: Left mandibular fragment with Mp FM 1505. Measurements: M, FM 1505: Length=30.0; Breadth=12.8; Trigonid length=26.1.

Description: The preserved part of the mandibular bone indicates a relatively low and narrow corpus below Mr The anterior border of the masseteric fossa ends below the posterior part of Mr The carnassial has typical felid features with very strong trigonid, wide angle between the cutting

93 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

blades and very reduced and labio-lingually compressed (“cutting“) talonid (Fig. 6). The labial surface is slightly convex and the lingual one straight to concave. The para- and protoconid are worn, having wearing facets on their labial surface and almost equal height. The protoconid is slightly longer and wider than the paraconid. Its cutting mesial edge is slightly longer than the cutting distal edge of the paraconid and both form a wide angle with a deep vertical groove in the lateral surface. A visible metaconid-talonid complex (where the metaconid is distinct as a small point just behind the distal ridge of the protoconid) is present at the distal end of the tooth. It is separated from the protoconid by a well expressed groove and has a cutting distal edge.

Comparison and discussion:

According to the revision of Bfaumont (1975) two machairodonts are known in late Miocene of Europe,Mackairodiis aphanistus in the Vallesian and Machairodus giganteus in the Turolian (the latter includes also the Chinese forms M. palanderi and M. tingii, which could be probably regarded as geographic subspecies). The recently described (Sotnikova 1999)

Fig. 7. Morphology of the lower carnassial inMachairodus from various Eurasiatic localities, a.M. apha nistus, FM 1505, Nessebar, Bulgaria, b. M. aphanistus, KTD-163, Kemiklitepe, Turkey(B onis 1994). c. M. cf. aphanistus MTN-3173, Montredon, France(B eaumont 1988). d. M. aphanistus, Cm 244, Charmoille (Beaumont 1975). e. M. aphanistus, TD 1177, F.ppelsheim, Germany (Bfaumoni 1975). f. “M. laskarevi“ (=M. aphanistus),rXGV\-1 /2257, Kalfa, Moldavia (Sotnikova 1992). g. “M. aphanistus“ (= M. giganteus), Mahmutgazi,Turkey (Sc hmidt -K ittler 1976). h. M. gi ganteas, Slq-936, Axios valley, Greece (A rambourg & P iveteau 1929). i.M. giganteus, Pikermi, Greece. (W agner , 1857). j. M. giganteas, Inst. Palaeont. Vienna, Pikermi, Greece (Beaumont 1975) k. M. giganteus, AMNH 20606, Samos, Greece(B eaumont 1975). 1. M. giganteus, Museum Berne, Samos, Greece (Beaumont 1975). m. M. cf. giganteus, Pavlodar (Bi aumont 1975). n. M. giganteus, Taraklia, (Bi aumont 1975). o. M. ex gr giganteus. Baccincllo, Italy (Rook et al. 1991). p.“M. kurteni“, Kalmakpai, Kazakstan(S otnikova 1992).

94 © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zobodat.at

M. kurteni is very close to M. giganteus, while M. laskarevi resembles to M. apbanistus. M. alberdiae from Spain is synonymized with M. apbanistus (Beaumont 1988). M. romeri from Turkey with a plesiomorphic lower carnassial and double mesial blade on the upper canine could be a separate species or a subspecies ofM. apbanistus (Sohmidt -K ittler 1976; Morlo 1997). The two species M. apbanistus and M. giganteus are well distinguished by a series of characters. Concerning the lower carnassial two morphotypes have been recognized: a. The first and more primitive has a relatively broad (cutting and crushing) trigonid and a well expressed metaconid-talonid complex (plesiomorphic). It is present M.in apbanistus of Eurasia (inch M. “laskarevi“ and M. “romeri“ as well as M. alberdiae from Spain). b. The other has a relatively narrow and cutting trigonid with a very reduced talonid, expressed as a basal prolongation of the cutting distal edge of the protoconid. The metaconid is not distinct from the talonid, nor from the distal surface of the protoconid (apomorphic). This morphotype is known in the later forms:M. giganteus of Eurasia (inch M. g. copei from Grebeniki) and M. kurteni from Kalmakpai - Kazakhstan. The first morphotype is quite clear in the type specimen from Eppelsheim, as well as in the material from Montredon and Charmoille (Fig. 7 c-e). In the Turolian M. giganteus the metaconid-talonid complex is quite reduced with variable development from rudimentary to absent. Even when it is present it seems like a distal cingulum (Fig. 7k-n). The size of the lower carnassial is also different in the two species. Although there is an overlapping (e.g. the studied material has the size of the Pikermi M. giganteus) the mean size of M. apbanistus is smaller than that of M. giganteus (Fig. 8). In this case the morphological characters of the tooth can easily distinguish the two species. The studied tooth from Nessebar has a large metaconid-talonid complex which is clear and well distinguished (Fig. 7a). In this feature it is very similar to the type material from Eppelsheim, as well as to that from Montredon (Fig. 7c,e).M. apbanistus was recently described from the locality of Kemiklitepe-D, in Turkey(B onis 1994). Even taking into account the bad preservation, the metaconid-talonid complex is important and large, like in the studied specimen. This locality is dated to early Turolian, MN 11 (Bonis et al. 1994) and represents an evidence for the presence of the species in Turolian. The comparison with the lower carnassial of the typical Turolian material distinguishes the studied tooth clearly. A piece of mandible with P4-Mj is known from the Turolian of Axios valley and referred as M. apbanistus (A rambourg & Pivf .TEAU 1929). Recently it was transferred

* M. aphanistus, Nessebar

♦ M. aphanistus, Eppelsheim

• M. aphanistus, Soblay A M. aphanistus, Charmoille

■ M. aphanistus, Mahmutgazi □ M. giganteus, Pikermi

O M. giganteus, Samos

O M. giganteus, Axios valley A M. ex gr. giganteus, Baccinello

<► M. kurteni, Kalmakpai, Kazakstan

Fig. 8. Scatter digramm comparing length / breadth of the lower carnassial in Alachairodus from various localities.

to M. giganteus (K oufos 2000a). The absence of the metaconid-talonid complex in this specimen separates it clearly from the studied M, (Fig. 7h). Several specimens ofM. giganteus are known from Pikermi and Samos (Fig. 7i-l). They preserve a metaconid-talonid complex which is more strongly reduced than that of the studied specimen. A mandibular fragment from Mahmutgazi (Turkey) has been described as M. aphanistus (Sohmidt -K ittler 1976). The size of the lower carnassial is somewhat larger than the typical M. aphanistus from Eppelsheim, taking a position between this species andM. giganteus (Fig. 8). The morphology of the tooth as it is shown in the illustration (Schmidt -K ittler 1976, pi. 4; fig. 5) is closer to M. giganteus. The metaconid-talonid complex is not large but it is like a vestigial distal projection of the cingulum (Fig. 7g). Based on this character the Mahmutgazi specimen seems to be closer toM. giganteus and distinct from the studied one. Some remains of a machairodont from Baccinello (Italy) are referred asM. ex gr. giganteus (Rook et al. 1991). The size of M, is very small comparatively to the other material of both M. aphanistus and M. giganteus (Fig. 8), while the metaconid-talonid complex is practically absent (Fig. 7o). In this feature it is closer to M. giganteus from Axios valley but the size is significantly smaller. As the late Miocene faunas of Italy are isolated this size reduction probably represents an adaptation ofM. giganteus to the local palaeoecological conditions. A small-sized machairodont, named M. alberdiae is known from the Vallesian of Spain (G jnsburg et al. 1981). Besides the smaller size it also differs from the studied specimen in having a relatively stronger metaconid. M. kurteni from the latest Turolian locality of Kalmakpai, Kazakhstan (C. Asia) has similar size of Mj (Fig. 8) but it differs strongly by the total lack of metaconid-talonid complex (Fig. 7p). It is very closeM. to giganteus representing the most evolved stage of the upper and lower carnassial of this species(S otnikova 1992). A mandible of a machairodont from Kalfa, Moldavia (upper MN 9) has been described under the name Ml laskarevi (Lungu 1978). It is very close toM. aphanistus both in size (Fig. 8) and morphology(S otnikova 1992). It preserves a small P, in the right branch which may be is an individual variability at the very early stage of M. aphanistus. The Mt is similar to the studied one (Fig. 7f).

Fig. 9. A. Logarithmic ratio diagram comparing the mandible of Nessebar hipparions. Standard: H. mediterranetim, Pikermi (K oufos 1987). 2. Muzzle length: middle of the line connecting the anterior borders of P, to a point situated between the two Ip 3. Premolar length (alveolar); 4. Molar length (alveolar); 5. Toothseries length (alveolar); 6. Distance from posterior end of M3-posterior border of the vertical ramus; 7. Muzzle breadth: breadth at the posterior borders of I3; 8. Height articular condyle - base of the horizontal ramus; 9. Height incisura mandibulae - base of the horizontal ramus; 10. Height of the jaw behing M3; 11. Idem between P4 and Mt; 12. Idem in front of P,; 13. Syntphysial length; 14. Minimal breadth of the symphysis; 16. Diastema P,-I3. ■ " NES-132 • NES-133 <0 H. primigenium, Ravin de la Pluic (K oufos 2000b). □ H. primigenium, Nikiti-1 (K oufos 2000c) ▲ H. primigenium, Vienna basin (Bf.rnor et al. 1988) A H. primigenium, Howenegg (Bfrnor et al. 1997) O H. brachypus, Pikermi (K oufos 1987).

B. Logarithmic ratio diagram comparing the metapodials of Nessebar hipparions. Standard: H. mediterranetim, Pikermi (K oufos 1987). □ Nessebar (Forsten 1972) A H. primigenium, Ravin de la Pluie (K oufos 2000b). <0 H. primigenium, Nikiti-1 (K oufos 2000c) O H. cf. depereti, Pentalophos 1 (K oufos 2000b).

The above-mentioned morphological and metrical comparisons of the studied tooth from Nessebar allow the determination to Machairodus aphanistus.

4. Biochronology

According to the Eastern Paratethys marine biostratigraphy the locality of Nessebar is considered as Sarmatian (= Vallesian). The fauna includes the following species: Hipparion mediterraneum, Hipparion nessebricum and Deinotherium giganteum (Bakalov & N ikolov

MANDIBLE

METAPODIALS

L 0.08 T

° 0.06 ■ • 9

d 0.04 -- ¡ f 0.02 ■ - f e o ■■ r

e - 0.02 ■ - n

C -0.04 - ■ e s -0.06

1962). Later on, another faunal list is given for Nessebar including the species: Hipparion nessebricum, Hipparionpraesulcatnm, Deinotherium bavaricnm and Sclnzocbocrusvallcsicnsis. The proposed age is again Sarmatian (= Vallesian), (N ikolov 1985). According to the description ofB akai.ov & N ikolov (1962) and N ikolov (1985) both hipparions from Nessebar are large-sized forms with rich enamel plication, the phs are narrow and deep and resemble to those of Vallesian forms. The characters of the Nessebar hipparions are: relatively massive metapodials, oval to pearshaped preorbital fossa situated far from the orbit, long and narrow protocone and rich enamel plication(P’o R S T iN 1978). These characters suggest a primitive hipparionine horse. In the wider Balkan peninsula the Vallesian is well known from Northern Greece, where five different localities are known. Three different hipparionine horses have been recognized, H. primigenium, //. aff. depcretiand H. macedonicum (K oufos 2000b, c). A comparison of the mandible and metapodials from Nessebar with those from Nothern Greece will give some more data about the Nessebar hipparions.H. macedonicum is a small-sized hipparionine horse and it is well distinguished by its size from those of Nessebar which are large forms. The available measurements for the Nessebar hipparions are few and they are given, those of the mandible by Bakalov & N ikolov (1962) and those of the metapo dials by FoR STEN (1978). The results are shown in the diagrams of Fig. 9. The mandible and metapodials from Nessebar are very close to those ofH. primigenium from “Ravin de la Pluie“ (RPl) and “Nikiti 1“ (NKT) while they differ from the other species H. aff. depereti (Fig. 9). This strong similarity is an indication for a Vallesian age of the Nessebar hipparions as both RPl and NKT are dated to late Vallesian (Bonis & K oufos 1999). As already mentioned, in the fauna of Nessebar Schizochoerusvallesiensis and “Deinotherium “ bavaricnm are also reported(N ikolov & T henius 1967; N ikolov 1985). Scbizocboerus vallesiensis is known from the Vallesian of Spain(FI unermann 1999) and it is one more indication about a Vallesian age for Nessebar. The presence ofProdeinotherium bavaricnm is another evidence for an old age, since it is known from early Miocene to the end of Vallesian (G ohlich 1999). The machairodont Al. apbanistus is known from the Vallesian (Beaumont 1975) and its presence in Nessebar gives an old feature to the fauna. Fiowever, it was also found in the locality “Kemiklitepe D “ (KTD) of Turkey which is dated to MN 11 (Bonis et al. 1994). Dinocrocnta gigantea is also known from the Vallesian in the Balkan region. It has been reported from the locality of “Pentalophos 1“ (PNT) of Axios valley dated to Vallesian (K oufos , 1995). It is also known from the Vallesian-Turolian of Turkey, China and Mongolia (Sc hmidt -K ittler 1976;IT owell & Petter 1985) and from the Vallesian of Moldavia(L ungu 1978). Despite the fact that the species is found not only in Vallesian but also in Turolian localities, the combination ofD. gigantea and M. apbanistus with Prodeinotherium bavaricnm and Scbizocboerus vallesiensis, as well as the similarities of the hipparions with//./»rimigenium from the Vallesian of Northern Greece are strong evidences for a Vallesian age of the Nessebar fauna. Nevertheless, it is difficult to recognize one of the two Vallesian MN zones because of the scarce material, the limited comparisons and the questionable determinations of the other faunal elements. The site of D. gigantea, near Blagoevgrad, is not precisely located and cannot certainly correlated with the known faunas and stratigraphy of the area. The Djerman Formation and the Barovska Formation outcrop in the northern part of the Blagoevgrad graben; both forma tions are dated to Maeotian-Pontian(M arinova 1993). A well known mammal locality, named Kocherinovo, is known within these deposits. According to unpublished data by I.N ikolov the following fauna has been determined: Aceratherium incisivum, Microstonyx major, aff. Helladotberium sp., Gazella sp., Indarctos sp., Elephantoidea indet., Orycteropus sp. The unpublished and lost Orycteropus skull from this locality resembles to the primitive forms of

the genus such as O. browni, O. pottieri, O. mauritanicus (determination of N. Spassov after existing photos). Such a fauna indicates that the locality could be at the Vallesian/Turolian boundary. It is worth noting here that the studied tooth keeps clear lignitic traces but no lignitic layers were found in the above mentioned formations. Some undivided Neogene deposits are also referred near the village of Dragodan, which possibly belong to the base of Neogene (Marinova 1993) and thus cannot include the studied tooth. Two areas near Blagoevgrad include lignites. The first is the area of Padaloto, west of Blagoevgrad but the deposits are considered as Oligocene (Zagorchev , pers. comm.). The other is the coal formation around Razlog which is the most probable for the studied tooth. The Neogene deposits near the villages of Oranovo and Arnautska mahala, north of Simitli, about 10 km southeast of Blagoevgrad (Fig. 1), include lignitic layers, while their age is most probably Sarmatian- Maeotian (i.e. roughly corresponding to Vallesian)(M arinova 1993).

Acknowledgements.

The second author thanks very much Prof. V. Fahlbusch and Prof. K. Heissig for providing to him access to collections at their disposal, as well as Dr S. Sen for providing him casts of Machairodus. Many thanks to Prof. E. Heizmann for reading the manuscripts and making usefull comments.

5. References

A ndrews , C. W. (1918): Note on some fossil mammals from Salonica and Imbros. Geol. Mag., 6:540-543, London. A rambourg , C. ( 1959): Vertébrés continentaux du Miocène supérieur de l’Afrique du Nord. Publications Service Carte Géologique Algérienne, n.s. Paléontologie, 4: 1-161, Algérie. A rambourg , C. & Piveteau , J. (1929): Les Vertébrés du Pontien de Salonique. Ann. de Paléont., 18: 59 138, Paris. Bakalov , P & N ikolov , L. (1962): Les fossiles de Bulgarie. X. Mammifères Tertiaires. Acad. Scie. Bulgarie, pp. 1-162, Sofia. Barishnikov , G. & A verianov , A. (1995): Deciduous teeth of carnivorous mammals (Order Carnivora). Part V. Fam. Protelidae and Hyaenidae. Proc. of the Zool. Inst. St. Petersburg, 263: 46-84. Beaumont , G. de (1975): Recherches sur les Félidés (Mammifères, Carnivores) du Pliocène inférieur des sables à Deinotherium des environs d’ Eppelsheim (Rheinhessen). Arch.Scie.Genève, 28: 369-405, Genève. Beaumont , G. de ( 1988): Contributions a l’étude du gisement Miocène supérieur de Montredon (Hérault). Les grands mammifères. 2. Carnivores. Palaeovertebrata, Mem. Extraordinaire 1:15-42, Montpellier. Bernor , R. L., K ovar -E dfr , J., Lipscomb, D., R ôgl , F., Si n , S. & T obien , H. (1988): Systematic, stratigraphie and palaeoenvironmental contexts of first-appearingHipparion in the Vienna basin, Austria. J.Vert. Palaeont., 8 (4): 427-452, N. York. Bi rnor , R. L., T obien , H., H ayek , L.-A. C. & M ittmann , H. W. (1997):Hippotherium primigemum (Equidae, Mammalia) from the late Miocene of Howenegg (Hegau, Germany). Andrias 10: 1-230, Karlsruhe. Bonis , de L. (1994): Les gisements de mammifères du Miocène de Kcmiklitepe, Turquie: 2. Carnivores. Bull. Mus. Nat. hist. Natur. Paris, 4e ser., 16, sect. C:19-39, Paris. Bonis , de L., Bouvrain , G., G eraads, D., K oufos , G. D., Sfn , S. & T assy , P. (1994): Les gisements de mammifères du Miocène supérieur de Kcmiklitepe (Turquie). 11. Biochronologie, paléoécologie et relations paléobiogeographiques. Bull. Mus. Nat. hist. Natur. Paris, 4e ser., 16, sect. C: 225-240, Paris. Bonis , de I.., Bouvrain , G. & K oufos , G. D. (1999): Palaeoenvironments of late Miocene primate localities in Macedonia, Greece. - in:A gusti , J., Rook, L. & A ndrfws , P. (eds): Hominoid evolution and climatic change in Europe, vol. I: The evolution of Neogene terrestrial ecosystems in Europe: 413-435, Cambridge (Cambridge University Press).

Bonis de L. & K oufos , G. D. (1999): The Miocene large mammal succession in Greece. - in:A gusti , J., Rook, L. & A ndrews , P. (eds.): Hominoid evolution and climatic change in Europe, vol. I: The evolution of Neogenc terrestrial ecosystems in Europe: 205-237, Cambridge (Cambridge Univ. Press). C hen , G. & Schmidt -K ittler, N. (1983): The deciduous dentition ofPercrocuta K retzoi and the diphyletic origin of Hyaenas (Carnivora, Mammalia). Paläont. Zeitschrift, 57: 159-169, Stuttgart. Forsten , A-M. (1978): A review of BulgarianHipparion (Mammalia, Perissodactyla). Geobios, 11:31-41, Lyon. G eraads , D. Sc G ülec, E. (1997): Relationships ofBarburofelispiveteaui (Ozansoy , 1965) a Late Miocene Nimravid (Carnivora, Mammalia) from Central Turkey. Journal of Vertebrate Paleontology, 17 (2): 370-375, N. York G insburg , L., M oralfs, J. Sc Soria, D. (1981): Nuevos datos sobre los carnivoros de los valles de Fuentidueha (Segovia). Estudios Geol., 37: 383-415, Madrid. G öhlich , U. B. (1999): The Miocene land mammals of Europe. 13. Order Proboscidea. - Rin:össner , G. E. Sc H eissig K. (eds): The Miocene land mammals of Europe: 157-168, München (Pfeil). H owf.ll, F. C. (1984): Observations on Carnivora from the upper Neogene Sahabi Formation (Libya). - in: Boaz, N. T., G aziry , A. W., H finzelin , J. df Sc E l-A rnauti , A. (eds): Neogene Palaeontology and Geology of Sahabi, Libya: 153-181, New York (Alan R. Liss Publishers). H owfll, F. C. Sc P etti r, G. (1985): Comparative observations on some middle and upper Miocene hyaenids. Genera: Percrocuta K retzoi, Allohyaena K retzoi, Adcrocuta K retzoi (Mammalia, Carnivora, Hyaenidae). Geobios, 18: 419-476, Lyon. FIünf .rmann , K. A. (1999): The Miocene land mammals of Europe. 20. Superfamily Suoidea.-in:Rössner , G. E. Sc H eissig , K. (eds): The Miocene land mammals of Europe: 209-216, München (Pfeil). K oufos , G. D. (1987): Study of the Pikermi hipparions. Part I: Generalities and taxonomy. Bull. Mus. nat. Hist, natur. Paris, 4e ser., 9, sect. C, (2): 197-252, Paris. K oufos , G. D. (1987): Study of the Pikermi hipparions. Part II: Comparisons and odontograms. Bull. Mus. nat. Hist, natur. Paris, 4e ser., 9, sect. C, (3): 327-363, Paris. K oufos , G. D. (1995): The late Miocene percrocutas of Macedonia (Greece). Palaeovertebrata, 24: 67-84, Montpellier. K oufos , G. D. (2000a): Revision of the late Miocene carnivores of the lower Axios valley. Mün. Geowiss. Abh., 39: 51-92, München. K oufos , G. D. (2000b): New material of Vallesian hipparions (Mammalia, Perissodactyla) from the lower Axios valley, Macedonia, Greece. Senckebergiana Lethaea, 80: 231-255, Frankfurt am Main. K oufos , G. D. (2000c) The hipparions of the late Miocene locality „Nikiti 1“, Chalkidiki, Macedonia, Greece. Rev. Palaeobiol., 19(1): 47-77,Geneve. K urt £n , B. (1957): Percrocuta K rftzoi (Mammalia, Carnivora), a group of Neogene hyaenas. Acta Zool. Cracov., 2: 375-404, Krakov. Lungu , A. N. (1978): Gipparionovaja fauna crednevo Sarmata Moldavii. (Hisnie miokopitaiosie). Jsdat. “Stiinza“, Kishinev, 131 p., Moldavia. M arinova , R. (1993): Explanatory note to the geological map of Bulgaria on scale 1:100 000. Blagoevgrad and Razlog sheets, pp. 37-67. In:Z agorvech , I. (ed.) Geological map of Bulgaria. Geol. and Geo physics Corp., Comm, of Geol. and Min. Res., Sofia. M elentis , J. K. (1967): Die Pikermifauna von Halmyropotamos (S. Euböa/Griechenland). Teil I: Odontologie. Ann. Geol. Pays Flell., 19: 283-411, Athens. M orlo, M. (1997): Die Raubtiere (Mammalia, Carnivora) aus dem Turolium von Dorn-Dürkheim 1 (Rheinhessen). Teil 1: Mustelida, Hyaenidae, Percrocutidae, Felidae. Courier Forsch.-Inst. Senckenberg, 197: 11-47, Frankfurt am Main. N ikolov , I. & T henius , E. (1967):Schizochoerus (Suidae, Mammalia) aus dem Pliozän von Bulgarien. Ann. Naturhist. Mus., 71: 329-340, Wien. N ikolov , I. (1985): Catalogue of the localities of Tertiary mammals in Bulgaria. Bulgarian Academy of Sciences. Palaeontology, Stratigraphy and Lithology, 21: 43-62, Sofia. Now (2002): Database of the Eurasian Neogene mammal faunas, http/Avww.helsinki.fi/science/now. Ozansoy , F. (1965): Etudes des gisements continentaux et des mammiferes du Cenozoique de lurquie. Mem. Soc. Geol. France, n.s. 44: 1-92, Paris.

Qiu, Z., Bj e J. & Y an , D. (1988): Discovery of the skull ofDinocrocuta gigantea. Vert. Palasiatica, 26: 128— 138, Peking (chinese with english summary). Qiu, Z., W u , W. & Qiu, Z. (1999): Miocene mammal faunal sequence of China: palaeogeography and Eurasian relationships. - in: R össner , G. E. & H eissig , K. (eds): The Miocene land mammals of Europe: 443-455, München (Pfeil). Pavlovic , M. ScT hf .nius , E. (1965): Eine neue Hyäne aus dem Miozän Jugoslawiens und ihreplylogenetische Stellung. Anzeiger der math.-naturw. Klasse der Österr. Akad. der Wiss., 9: 177-185, Wien. Rook, L., F iccarelli, G. & T orrf, D. (1991): Messinian carnivores from Italy. Boll. Soc. Paleont. Italiana, 30 (1): 7-22, Modena. Schlosser , M. (1903): Die fossilen Säugetiere Chinas nebst einer Odontographieder recenten Antilopen. Abh. Bayer. Akad. Wiss., 7: 372-464, München. Schmidt -K ittler, N. (1976): Raubtiere aus dem Jungtertiär Kleinasiens. Palaeontographica, Abt. A, 155: 1-131, Stuttgart. Soria, D. (1980): “Percrocuta“ y “Adcrocuta“ (Hyaenidae, Mammalia) en el Mioceno superior del area de Teruel. Estudios Geol., 36: 143-161, Madrid. Sotnikova , M. (1992): A new species ofMachairodus from the late Miocene Kalmakpai locality in eastern Kazakhstan (USSR). Ann. Zool. Fenn., 28: 361 -369, Helsinki. Spassov , N. (2000): The Turolian Hipparion-fauna and the character of the environment in the late Miocene of South-East Bulgaria. Rev. Bulg. Geol. Soc., 61 (1-3): 47-59, Sofia. W erdelin , L. (1996): Carnivores, exclusive of Hyaenidae, from the Later Miocene of Europe and Western Asia. - in: Bf.rnor , R., RirTSCHEL, S. & M ittmann , W. (eds): The Evolution of Western Eurasian Miocene Mammal Faunas: 271-289, New York (Columbia Univ. Press) W erdelin , L. & Solounias , N. (1991): The Hyaenidae: taxonomy, systematics and evolution. Fossils and Strata, 30: 1-104, Oslo. W erdelin , L. & Solounias , N. (1996): The Evolutionary Historyof hyaenas in Europe and Western Asia During the Miocene. - in: Bernor , R., Rietschel , S. & M ittmann , W. (eds): The Evolution of Western Eurasian Miocene Mammal Faunas: 290-306, New York (Columbia Univ. Press).

101